I am a machine learning researcher and engineer working at the intersection of mathematical optimization, scalable learning systems, and neural architecture design.

My primary interests include distributed and federated learning, communication-efficient optimization, efficient LLM training, and Mixture-of-Experts architectures. I am particularly interested in how learning algorithms can adapt to limited compute, heterogeneous infrastructure, and complex interactions between models or experts.

My background in mathematics, physics, and computer science shapes the way I approach machine learning research. I aim to understand not only whether a method works, but why it works, how it behaves under different constraints, and how its theoretical properties translate into real training systems.

I enjoy working across the full research cycle, from studying mathematical foundations and formulating hypotheses to implementing algorithms in PyTorch, building distributed experimentation pipelines, running ablation studies, and analyzing convergence, efficiency, and model quality. I am most motivated by projects that connect principled ideas with rigorous experiments and reproducible engineering.

More broadly, I am interested in autonomous research agents, AutoML, retrieval-augmented systems, representation learning, and new approaches to organizing computation inside neural networks.

I am open to research collaborations, internships, and ambitious projects in optimization, distributed machine learning, and large-scale AI systems.